Crop Science Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 August 2005
Published in Crop Sci 45:1809-1816 (2005)
© 2005 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in Crop Science
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Caetano-Anollés, G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Caetano-Anollés, G.
Agricola
Right arrow Articles by Caetano-Anollés, G.
Related Collections
Right arrow Cell Biology & Molecular Genetics
Right arrow Crop Genetics
Right arrow Evolution

CROP BREEDING, GENETICS & CYTOLOGY

Evolution of Genome Size in the Grasses

Gustavo Caetano-Anollés*

Dep. of Crop Sciences, Univ. of Illinois at Urbana-Champaign, 332 NSRC, 1101 W. Peabody Drive, Urbana, IL 61801

* Corresponding author (gca{at}uiuc.edu)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The grasses (Poaceae) exhibit marked conservation of gene content and order (synteny and colinearity), a feature that promises the extension of genomic information from one grass species to another using a unified comparative approach. Grasses also show striking differences in the size of their genomes. Since this complicates the treatment of the grass family as a model genetic system, there is need to better understand the patterns and processes that drive genome size evolution. This requires knowledge of phylogenetic relationships, especially deep branching patterns that unify the grass subfamilies. In this study, a phylogeny of 66 grass species with known genome size, most of them diploid, was assembled. The phylogeny described relationships integrated from shared and derived characteristics in molecular and morphological data and the branching order of basal lineages recently inferred from RNA structure and large-scale chromosomal rearrangements. Evolutionary changes in genome size that exclude the effect of polyploidization were traced along the branches of the tree using parsimony methods of character state reconstruction. Most levels of change did not exceed twofold, and few exceeded threefold. The frequency of changes in genome size appears to decrease exponentially with the magnitude of change. The ratio of increases-to-decreases in genome size increased in the order Ehrhartoideae, PACCAD (panicoids, arundinoids, chloridoids, centothecoids, aristidoids, and danthonioids), and Pooideae clades. However, there were clear patterns of increase and decrease in all major clades, and notable genome size changes in the Pooideae and Chloridoideae subfamilies. This shows different tendencies in genome size diversification in these major grass lineages. Depending on the tracing method, the genome of the ancestor of the grass family had 3.0 to 5.2 pg DNA per 2C nucleus. Results extend early proposals that suggest genome size has both increased and decreased along grass lineages, and show that different models of character evolution imparted different frequencies and levels of change along the branches of the trees.

Abbreviations: C-value, the DNA amount in the unreplicated haploid nucleus of an organism and stands for constant • GPWG, Grass Phylogeny Working Group • ITS, internal transcribed spacer • MYA, million years ago • PACCAD, panicoids, arundinoids, chloridoids, centothecoids, aristidoids, and danthonioids • rRNA, ribosomal RNA • SP, squared-change parsimony • SRP, signal recognition particle • WP, Wagner parsimony


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
THE GRASS FAMILY (Poaceae) is a diverse and ecologically dominant group of monocotyledonous plants. About 10000 species have been described to date, and from a taxonomical point of view, these have been categorized into 700 grass genera using architectural, anatomical, and embryological characteristics (Clayton and Renvoize, 1986; Watson and Dallwitz, 1992). Genera have been unified into subtribes, tribes, supertribes, and subfamilies. This makes the grasses one of the largest families of flowering plants. Grass species share a unique floral and inflorescence structure, the spikelet, which appears to have originated in steps throughout evolutionary history (Kellogg, 2001). They also produce a distinctive fruit structure (the grain or caryopsis) that is unique within flowering plants and has immense agricultural value. In fact, the grasses include all major cereals and most minor grains, and these have been the staple of human kind since their domestication 10000 years ago.

Grass species differ widely in morphology and physiology. Similarly, they show striking differences in the size of their genomes, with DNA content ranging from 0.5 to 40 pg DNA per 2C nucleus (Bennett et al., 1982, 2000; Bennett and Smith, 1976, 1991; Bennett and Leitch, 1995, 2003). Besides polyploidization and duplication, genome size differences are embodied in noncoding and repetitive DNA (SanMiguel et al., 1996, 1998). These differences often result from mutational mechanisms of nucleic acid addition and loss, such as transposition (transposable element activity), spontaneous insertions and deletions, and chromosomal rearrangements (Petrov, 2001). For example, genome expansion in Arabidopsis appears to be counteracted by genome reduction through illegitimate recombination (Devos et al., 2002). Genome size differences are important in the context of conservation of gene content and order (synteny and colinearity) and the possibility of extending genomic information directly from one grass species to another using comparative genomic approaches (Devos and Gale, 2000). This has motivated evolutionary studies that trace changes in genome size throughout the history of diversification of the grass family (Bennetzen and Kellogg, 1997; Kellogg, 1998). While these phylogenetic studies have shown that increases in genome size have occurred over evolutionary time, it is not clear if decreases are equally likely (Bennetzen and Kellogg, 1997). Furthermore, establishing the direction of genome size evolution is difficult and often intimately linked to phylogenetic inferences and resolution of deep branches in grass phylogenies.

In this study, well-established methods of character state reconstruction were used to trace evolutionary changes in genome size along the branches of a phylogenetic tree that describes the evolution of major grass lineages. The phylogeny that was used includes grass species with diploid genomes of known DNA content and integrates results from a recent and comprehensive phylogenetic study based on macromorphology, anatomy, biochemistry, and the sequence of chloroplast and nuclear genes (Grass Phylogeny Working Group, 2001] with inferences from RNA structure and large-scale chromosomal rearrangements (Caetano-Anollés, 2005). Different models of character change were used to reconstruct evolution of genome size and evaluate patterns of genome size increase and reduction in these grasses.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Plant C-Values and Selection of Grass Species
A representative group of grass species for which there are both phylogenetic and genome size information was selected. DNA amounts in unreplicated diploid nuclei (2C-values; a C-value is the DNA amount in the unreplicated haploid nucleus of an organism) were obtained from the Plant DNA C-values Database at http://www.rbgkew.org.uk/cval/homepage.html (verified 15 Apr. 2005; Bennett and Leitch, 2003). This release encompasses recently described plant C-values (Bennett et al., 2000). With few exceptions, only grass species with diploid chromosome numbers were included in the analysis to avoid complication due to polyploidy, and when available for a genus, species with minimum and maximum 2C-values were selected, to bracket genome size instead of targeting a median 2C-value for the genus. Tetraploid species representative of Pooideae (Brachypodium pinnatum (L) P. Beauv., Brachypodium sylvaticum (Hudson) P. Beauv., and Glyceria fluitans (L.) R. Br.), Danthonioideae (Danthonia decumbens (L.) DC.), Chloridoideae (Eragrostid tef (Zucc.) Trott.), and Arundinoideae (Molinia caerulea (L.) Moench.) were included in the analysis to balance sections of the tree. Their genome size was expressed as 2C-values using normalization factors. Factors ranged 0.51 to 0.53 in value and were inferred by comparing genome sizes of diploid and tetraploid plants within each genus in the database. However, the focus was on diploid species in an attempt to simplify the treatment of genome size increases of relatively recent origin that are driven by polyploidization.

Phylogenetic Assumptions
A phylogeny depicting the history of diversification of major lineages of the grass family was assembled from the diploid species selected, and genome size was traced along its branches. Phylogenetic relationships of diploid grasses summarized by Kellogg (1998) were superimposed on segments of the skeletal phylogeny proposed by the Grass Phylogeny Working Group (2001) rerooted in the Ehrhartoideae (Caetano-Anollés, 2005). The Grass Phylogeny Working Group research consortium described the evolutionary relationships of representative species within major grass subfamilies using macromorphology, anatomy, biochemistry, and molecular features such as restriction endonuclease maps of the chloroplast genome and the nucleotide sequence of chloroplast (ndhF, rbcL and rpoC2) and nuclear (phyB, and waxy) genes and intergenic ribosomal RNA (rRNA) spacers, and maximum parsimony methods of tree reconstruction (Grass Phylogeny Working Group, 2001). The Grass Phylogeny Working Group phylogeny combined 8 character sets, some of which support strongly (chloroplast restriction-morphological data), moderately (rbcL), and weakly [chloroplast restriction data and internal transcribed spacer (ITS) rRNA] the existence of a Pooideae and a PACCAD clade. Other character sets were inconclusive or supported other groupings that were slightly more parsimonious, but none included any morphological synapomorphies. Deep branching patterns that reroot the tree in the Ehrhartoideae and support a sister clade relationship between the Pooideae and PACCAD clade were derived directly from geometrical and statistical features describing the structure of signal recognition particle (SRP) RNA, the small subunit of rRNA, enod40 mRNA and ITS rRNA, and large-scale chromosomal rearrangements (insertions, translocations, and instances of chromosomal orthology) using maximum parsimony methods (Caetano-Anollés, 2005). For RNA, molecules were characterized by attributes such as nucleotide length of molecular components, thermodynamic properties such as minimum Gibbs free energy increments, or statistical parameters that describe the stability and uniqueness of folded conformations. Attributes were then treated as linearly ordered multistate characters that were polarized by a model of character state transformation in which structures with increased molecular order and minimum frustration were defined as being ancestral (Caetano-Anollés, 2002a, 2002b). This approach is supported by considerations in statistical mechanics, produces intrinsically rooted trees that "embed structure and function directly into phylogenetic analysis" (Pollock, 2003), and was used successfully to reconstruct a phylogeny of the living world (Caetano-Anollés, 2002a) and study ribosomal evolution (Caetano-Anollés, 2002b). In the analysis of chromosomal rearrangements, phylogenetic reconstruction supports the proposal that genetic linkage blocks that are freestanding in rice were ancestral and that rearrangements that result in translocations, insertions, and redistribution of rice linkage blocks were derived events (Gale and Devos 1998).

Evolutionary Tracing of Genome Size in the Grass Family
Genome size was traced along the individual branches of the phylogenetic tree. Ancestral character states were reconstructed using algorithms for squared-change (Maddison, 1991) and Wagner parsimony (Swofford and Maddison, 1987) in MACCLADE v. 3.08 (Maddison and Maddison, 1999). Squared-change parsimony (SP) minimizes the sum of the squared changes on the branches of the tree and can be considered a Bayesian probability estimate under a Brownian motion model of evolution. Wagner (linear) parsimony (WP) minimizes the sum of the absolute value of changes on the branches of the trees. Because this produces a range of equally parsimonious values, only minimum values were chosen.


   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Phylogenetic analysis of the structure of functional RNA and a cladistic study of large-scale chromosomal rearrangements supported an early branching of the Ehrhartoideae subfamily (Caetano-Anollés, 2005). This information, together with the skeletal phylogeny compiled by the Grass Phylogeny Working Group (2001) (see Materials and Methods), was used to define an extended phylogeny of 66 grass species with known DNA content (Fig. 1 and 2). Many grass species present in the Grass Phylogeny Working Group (GPWG) phylogeny could not be included because their C-values had not been established (including outgroup taxa such as Anomochloa, Pharus, and Puelia). The assembled phylogeny was used to predict the genome size of hypothetical ancestors (nodes) along evolutionary lineages, using the agnostic and increase-only models proposed by Bennetzen and Kellogg (1997). The agnostic model considers that both increases and decreases in genome size are equally likely, and was implemented here using squared-change (SP) and linear (WP) parsimony methods of ancestral character reconstruction. Squared-change parsimony revealed 60 increases, 66 decreases, and three instances of no change in genome size along the branches of the tree (Fig. 1). The WP showed 42 increases, 21 decreases, and 66 instances of no change (Fig. 2). The increase-only model considers that the genome size of an ancestor cannot be larger than the smallest found in its descendants. Under this unidirectional model, there were 62 increases in genome size along the branches of the tree (data not shown). Figure 3 compares the number and levels of change along the phylogeny of the grasses obtained with the three ancestral character state reconstruction methods, and Table 1 shows several statistics derived from these analyses. The frequency of changes in genome size appears to decrease exponentially with the magnitude of change (Fig. 3). The fraction of changes exhibiting more than twofold increases or decreases was minimum when using SP (13.5%) but increased with WP (28.5%) and the unidirectional model (56.5%). Higher levels of change (> threefold) occurred with less frequency and in the same order; 4, 14.3, and 37.1% for the SP, WP, and unidirectional models, respectively. Moreover, these same trends were observed when considering only increases or decreases in genome size or when averaging levels of change. The agnostic models produced larger total averages (1.5- and 1.9-fold) than the unidirectional model (fourfold) (Table 1).



View larger version (76K):
[in this window]
[in a new window]
  		Fig. 1. The evolution of genome size in the grasses inferred using the squared-change parsimony (SP) criterion. Ancestral genome sizes (in pg DNA per 2C nucleus) were reconstructed as continuous-valued characters on a cladogram of diploid grass species using SP with the rooted option. The phylogeny of the grasses was rooted using phylogenetic information derived from RNA structure. Closed and open circles indicate nodes with more than twofold increases or decreases in genome size, respectively. The node defining Bromeae leads to Bromus species and the node defining Triticeae leads to Triticum and Aegilops species.

 


View larger version (75K):
[in this window]
[in a new window]
  		Fig. 2. The evolution of genome size in the grasses inferred using Wagner's linear parsimony criterion. Ancestral genome sizes (in pg DNA per 2C nucleus) were reconstructed as continuous-valued characters on a cladogram of diploid grass species using linear parsimony with minimum equally parsimonious values. The phylogeny of the grasses was rooted using phylogenetic information derived from RNA structure. Closed and open circles indicate nodes with more than twofold increases or decreases in genome size, respectively.

 


View larger version (27K):
[in this window]
[in a new window]
  		Fig. 3. Frequency histograms showing the frequency distribution of changes in genome size along branches of the phylogenetic tree of diploid grass species, inferred using agnostic (squared-change and Wagner parsimony) and unidirectional models. The levels of genome size change are given as fold increases or decreases occurring at individual nodes. Values are binned in class intervals of width 0.2.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Statistics describing changes in genome size on the phylogenetic tree of diploid grasses.{dagger}

 
The genome size of the ancestor of the grasses was inferred to be 0.5 pg (unidirectional model), 3 pg (WP), or 5.2 pg (SP) of DNA per 2C nucleus. This implies that genome size increased five (SP), nine (WP), or 55 (unidirectional model) times throughout the history of the grass family, and under the agnostic models, decreased six (WP) or 10 (SP) times (Table 1).

Genome size increased in the phylogenetic tree of the grasses under all models. Genome size values reconstructed at internal nodes showed clear patterns of increase (cf. nodes defining the ancestor of grasses, Pooideae, Bromeae, and Triticeae) and decrease (e.g., in Ehrhartoideae and the PACCAD clade). There were notable genome size changes in the Pooideae [ancestors of Lygeum spartum Loefl. ex L. (two- to sevenfold increase), Agropyron (19-fold increase and 10- to 12-fold decrease), and Brachypodium (four to fivefold decrease)] and Chloridoideae [ancestors of Eragrostis tef (Zucc.) Trott. (fourfold decrease) and Spartina anglica C.E. Hubbard (four- to 22-fold increase)]. Overall patterns of genome size increase were also evident when comparing the number of increases and decreases occurring in individual clades [as increase/decrease ratios (r); Table 1]. In all three models, ratios increased in the order Ehrhartoideae, PACCAD, and Pooideae clades, showing different tendencies in genome size diversification in these major plant groups.


   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Classical and molecular phylogenetic analysis has shown that the grasses arose about 55 to 70 million years ago (MYA), probably as a single monophyletic lineage (Jacobs et al., 1999). However, the exact order of grass diversification events has been difficult to establish, and remains contentious (Kellogg, 1998, 2001). In a recent study, a consortium of researchers described the evolutionary relationships of representative species within major grass subfamilies (Grass Phylogeny Working Group, 2001). A well-supported consensus tree was inferred from 2143 informative morphological and molecular characters using parsimony. The phylogenetic tree showed the basal diversification of three subfamilies, the Anomochloideae, Pharoideae, and Puelioideae, one early major radiation comprising the Bambusoideae, Ehrhartoideae, and Pooideae subfamilies, followed by the PACCAD clade with all C4 lineages and several C3 plants (the clade includes the Panicoideae and Chloridoideae subfamilies). However, this tree lacked phylogenetic resolution at the base of the Poaceae, and subfamily relationships could not be unambiguously established. In a recent study, phylogenetic relationships inferred directly from the structure of functional RNA molecules (SRP RNA, enod40 mRNA, the small rRNA subunit, and ITS spacers) and large-scale chromosomal rearrangements (derived from comparative genetic mapping) were used to establish an order for the diversification of major grass lineages and a sister relationship of the Pooideae and the PACCAD clade (Caetano-Anollés, 2005). The advantage of using elements of RNA structure and large-scale chromosomal rearrangements is that these are likely to be more stable through evolutionary time than nucleotide sequences or microsyntenic relationships, which will be adversely affected by multiple substitutions or rearrangements, respectively. The method is therefore appropriately suited to uncover deep phylogenetic relationships.

A correct phylogeny of the grasses defines a direction of change and is critical in our efforts to establish tendencies in genome size evolution (Bennetzen and Kellogg, 1997). In fact, alternative rooting of the major grass subfamilies affect inferences of ancestral genome size when DNA contents (C-values) were traced on a phylogeny of the grasses using SP (see below). Given the deep-branching relationships inferred using the novel comparative approach (Caetano-Anollés, 2005) and the detailed genetic relationships of the Grass Phylogeny Working Group skeletal phylogeny (Grass Phylogeny Working Group, 2001), I assembled a phylogenetic tree of representative grass species and traced the evolution of genome size along its branches using different models of character change (Fig. 1–3, Table 1). Two kinds of algorithms were used to find ancestral states that were most parsimonious. One minimizes the sum of the absolute values of the changes using a linear Wagner parsimony criterion (Swofford and Maddison, 1987), and the other minimizes the sum of the squared changes using a Brownian motion model of evolution (Maddison, 1991). Using these parsimony criteria, genome size increased and decreased along the lineages of the grass family at varying levels. Alternatively, a unidirectional model that prohibits genome size decreases was used and resulted in higher levels of genome size increase along the tree. The existence of increase-only mechanisms is an unrealistic evolutionary scenario. Genome size reduction explains the excess of sequences flanking BARE-1 retrotransposons in barley as remnants of reductive recombination events (Vicient et al., 1999; Shirasu et al., 2000), and DNA sequencing of a large 211-kb DNA segment in diploid wheat revealed a complex pattern of genome rearrangement, including deletion of large DNA fragments containing retroelements (Wicker et al., 2001). Furthermore, differences in rates of DNA loss appear to be important determinants of genome size evolution in insects (Petrov et al., 2000), and could play a similar role in the grasses. Similarly, decrease-only explanations for genome size evolution are highly unlikely. There is ample evidence that grass genomes have increased in size by chromosomal duplication and the effects of repetitive DNA (SanMiguel et al., 1996, 1998; Bennetzen, 2000; Gaut et al., 2000). For example, the maize genome increased considerably in size by a retrotransposon invasion that began {approx}5.2 MYA (SanMiguel et al., 1998). Similarly, recent and hidden polyploidization events appear to be a widespread phenomenon in the grasses, and both result in genome size increases (Levy and Feldman, 2002).

While recent polyploidy can be easily recognized, many species currently regarded as bona fide diploids could actually represent paleopolyploids, ancient polyploids with disomic inheritance and progenitors that cannot be identified using cytology or DNA markers. In fact, polyploidy could have occurred in the lineage of at least 70% of angiosperms (Masterson, 1994) and appears a revolutionary and ongoing process in the grasses (Levy and Feldman, 2002). The controversial proposal that genome evolution is mainly driven by whole-genome duplication (Ohno, 1970) has been recently used to explain chromosomal and synteny patterns in plants (e.g., angiosperms; Bowers et al., 2003) and fungi (e.g., hemiascomycete fungi; Dujon et al., 2004). Even vertebrates are believed to have experienced two rounds of paleopolyploidization (Wolfe, 2001). However, evidence in favor of paleoploidization events has been criticized as observations could be more parsimoniously explained by local duplication and genomic rearrangements (e.g., Hughes and Friedman, 2003). For example, gene interleaving patterns of synteny in yeast (Saccharomyces cerevisiae) were better explained by segmental duplication and recombination when analyzing the gene complements of entire chromosomes with advanced rearrangement algorithms (N. Martin et al., 2004, unpublished data). Phylogenetic analysis and comparative genomics will ultimately help identify the role of ancient polyploidy in evolution of genome size, as C-values are traced along branches delimiting phylogenetic hypotheses.

Character tracing offers the opportunity to study how changes in genome size distribute along different grass lineages and draw inferences on the genome size of hypothetical ancestors by assigning character states to internal nodes of the trees. Most changes in grass genome size did not exceed twofold increases or decreases, and few exceeded threefold levels of change. This occurred when using the two parsimony models of character evolution. There were clear patterns of increase and decrease in the Ehrhartoideae, Pooideae, and PACCAD clades, and notable genome size changes in the Pooideae and Chloridoideae. There were also notable increases and decreases in genome size occurring within an individual genus (e.g., Agropyron). In fact, this was expected. For example, a clear reduction of genome size has been recently proposed in one of two distinct lineages of Sorghum (Price et al., 2005). Parsimony analysis suggested an ancestor of the grass family with a genome size between 3.0 and 5.2 pg DNA per 2C nucleus (2.9 x 109 and 5.1 x 109 base pairs). This represents a genome that is six to 10 times bigger than the smallest (Oropetium thomaeum Trin.) and five to seven times smaller than the largest (Lygeum spartum) diploid genome described (Bennett and Leitch, 2003). Interestingly, the analysis of a phylogeny of 37 diploid species using the SP method resulted in an ancestral genome size (3.5 pg DNA per 2C nucleus) that was much lower than the estimate obtained in this study (Kellogg, 1998).

Only a few studies have traced genome size evolution along the branches of phylogenetic trees using character reconstruction methods. This is because the exercise requires both robust phylogenetic hypotheses and comprehensive sampling of DNA content among clades, conditions that only recently are beginning to be met. Recent studies suggest that the genome size of ancestors of angiosperms (Soltis et al., 2003) and land plants (Embryophyta) (Leitch et al., 2005) was small (≤1.4 pg DNA per 1C nucleus). Most of the major clades within angiosperms (e.g., monocots, magnoliids, eudicots) appear to have also small ancestral genomes, showing many possible instances of genome size increase and decrease in clades that occupy derived positions of the trees (Soltis et al., 2003; Leitch et al., 2005). Within this framework, monocotyledoneous plants appear to exhibit several independent instances of genome size increase in major lineages, but this depends fundamentally on how well phylogenies resolve in each lineage (Leitch et al., 2005). In this regard, the present study confirms an instance of overall genome size increase in a derived lineage of the Commelinids. Clearly, a comprehensive character reconstruction effort will be needed to obtain more accurate estimates of ancestral genome size for internal nodes along the lineages of terrestrial plants and a better picture of genome size evolution.

While the WP method may be too conservative, the SP method may overestimate the ancestral size of a genome. The validity of hypotheses of character change was therefore tested with an example of genome evolution in the Panicoideae. Pennisetum and maize diverged about 29 MYA, followed 9 million years later by the divergence of the two diploid progenitors of maize (Gaut et al., 2000). Genome size of the progenitors of Zea and Tripsacum (that diverged {approx}4.7 MYA) doubled as the result of a segmental allotetraploid event, which occurred sometime between the divergence of sorghum ({approx}6.5 MYA) and the rediploidization of the ancestors of maize ({approx}11 MYA). This allotetraploid event was followed by genome rearrangement and then by a retrotrasposon invasion that began before the split of Zea and Tripsacum (Gaut et al., 2000). Comparison of the retrotransposon-invaded Adh1 region of maize and the retrotransposon-free Adh1 region in barley suggested that the genome size of the ancestor of maize doubled in size (SanMiguel et al., 1998). The genome size of hypothetical ancestors of maize that are common to Pennisetum, Sorghum, and Tripsacum, was 5.3, 6, and 6.9 pg per 2C nucleus when inferred using the SP method (Fig. 1) and 3, 3, and 5.3 pg per 2C nucleus when inferred using the WP method (Fig. 2). Only the WP method accounted for the expected retrotransposon-induced doubling of genome size proposed during the late history of diversification of maize. None of the two methods revealed a major increase resulting from segmental allotetraploidy. The SP method showed only modest increases (15–30%) in genome size of ancestors along the maize lineage (starting {approx}29 MYA). In fact, these increases were lower than those observed occurring within species of Zea (30–72%). Only the unidirectional character tracing method accounted for increases due to allotetraploidy and retrotransposon proliferation, but this is an unrealistic model that discards mechanisms of genome contraction proposed linked to these same two phenomena.

The results of this study extend early proposals that suggest genome size has both increased and decreased along phylogenetic lineages of the grass family (Bennetzen and Kellogg, 1997). Bidirectional evolution of plant genome size appears a widespread phenomenon. It was recently reported in the cotton tribe Gossypeae (Wendel et al., 2002) and its dynamic nature described in angiosperms (Soltis et al., 2003) and land plants (Leitch et al., 2005). The present study also shows that different models of character evolution imparted different frequencies and levels of change along the branches of the trees. While the SP method favored decreases over increases, the linear WP method minimized the sum of the absolute values of change along branches of the grass tree and produced a more even distribution of genome size. Note that the SP method maximizes the Bayesian posterior probability (probability of an hypothesis given the data) of reconstruction of character states at ancestral nodes when changes are inversely weighted by the length of the branches (Maddison, 1991). However, the method may overestimate changes in the trees. Future efforts to mitigate difficulties in the prediction of genome size of hypothetical ancestors may involve weighting schemes that take into account variation in rates of genome size evolution along branches of the tree.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Research was partially sponsored by the Dep. of Biology, Univ. of Oslo, Norway and Vital NRG, Knoxville, TN 37919, and supported by Vital NRG, the International Atomic Energy Agency in Vienna (FAO/IAEA RCP580-8151), CSRESS-USDA-ILLU Hatch 802-316, and NSF MCB-0343126. Any opinions, findings, and conclusions expressed in this material are those of the author and do not necessarily reflect the views of the funding agencies. A preliminary version of this work was presented at the 6th International Congress of Plant Molecular Biology, Quebéc, Canada (18–24 June 2000).

Received for publication October 13, 2004.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 


Related articles in Crop Science:

THIS ISSUE IN CROP SCIENCE

Crop Science 2005 45: x. [Full Text]  



This article has been cited by other articles:


Home page
 ANN BOT (LOND)Home page
P. Smarda, P. Bures, L. Horova, B. Foggi, and G. Rossi
Genome Size and GC Content Evolution of Festuca: Ancestral Expansion and Subsequent Reduction
Ann. Bot., February 1, 2008; 101(3): 421 - 433.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
C. Vitte and J. L. Bennetzen
Eukaryotic Transposable Elements and Genome Evolution Special Feature: Analysis of retrotransposon structural diversity uncovers properties and propensities in angiosperm genome evolution
PNAS, November 21, 2006; 103(47): 17638 - 17643.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in Crop Science
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Caetano-Anollés, G.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Caetano-Anollés, G.
Agricola
Right arrow Articles by Caetano-Anollés, G.
Related Collections
Right arrow Cell Biology & Molecular Genetics
Right arrow Crop Genetics
Right arrow Evolution

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Plant Registrations Soil Science Society of America Journal
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality